A complete tutorial for training a YOLOv8n detection model from scratch on NVIDIA CUDA and compiling it into a Hailo HEF for deployment on NE503, covering dataset preparation, training, static-shape ONNX export, Hailo DFC quantization compilation, and device deployment.
Build, deploy, and run your first container application on NE503 from scratch — the complete closed loop from writing code to Web Console verification.
This document details the model training and deployment workflow for the NE301 series AI smart camera. It covers setting up a YOLO training environment, model training, exporting to TFLite format, model quantization and evaluation, and finally deploying the quantized model to the NE301 device using either Makefile build or Web UI upload methods, helping developers achieve efficient edge AI model deployment.
Deploy a real AI inference application Person Detection on NE503, mastering Python SDK calls, model and video stream discovery, permission configuration, and Web console verification.